(FIG-20)
Well I have to show you one more complicated figure -the only one complicated figure. The reason I'm showing this is： do we have any other good seismic evidence that indicates that these processes are taking place ? Well, in order to show that, I need to show this. This is the fault plane, and rupture starts from here and propagates at some velocity v; in the simplest case, on the left, the friction drops suddenly almost as a step function. When it slips the friction drops, and in this case the displacement goes something like this, and this is the final displacement on the fault plane. In the other case, friction doesn't drop abruptly.
In the case of small earthquakes, as I said, because thermal process doesn't take over, friction can drop very gradually. In this case, the displacement goes rather gradually like this, so you have two different types of earthquakes. In one case it's very brittle： friction drops, so that to get up to this displacement, things go very fast. So this is more or less the situation for large earthquakes - magnitude 6 or larger, and this is the situation for small earthquakes： magnitude 4 or maybe smaller.
So if you take the ratio of velocity to total displacement, you should be able to see the difference between small earthquakes and big earthquakes. As I said, for small earthquakes this ratio should be small, because fault motion may go relatively slowly. For large events, bemuse of lubrication, fault motion should be fast, so that this ratio must be large.
(FIG-21)
So the easiest way to see this difference is to compute this ratio from observations of radiated energy, ER, which is proportional to velocity, and the seismic moment, M0, which is proportional to displacement. Seismologists have been measuring energy since the 1940's. Gutenberg started this one. We haven't made a big progress until recently, and the reason is that to determine seismic energy is extremely difficult. However, recently this situation has improved tremendously. This is the magnitude from 1,2, 3, 4, 5, 6, ....and these are the data of ER and MO. This shows the ratio, and for large events 5, 6, 7, the ratios are all here. Now absolute values are not terribly important. For small events the ratio is over here, and this difference is almost a factor of 100.
This kind of data has been around in seismology, but no one paid attention to it. The reason is that everyone was very suspicious about the quality of data. And there is again some important improvement. These data for small events came from a down-hole instrument at Cajon Pass drilling site, and they're very accurate. And these data for large events are from California where recently we had a 100-station broad-band network, so we can improve the quality of measurements. So this factor of 100 difference in this ratio between small and large earthquakes basically reflects the difference in the condition in the fault plane.
(FIG-22)
I'm just showing you the same diagram in a more schematic way. If you take the ratio, ER/M0, roughly equivalent to velocity/displacement ratio, the ratio for large events is here, and the ratio for small events is here, and the difference is about a factor of 100. So this is a sort of seismological evidence that there is transition from small to large earthquakes which involves drastic change in the dynamics of faulting. Of course in the actual fault zone, the structure can be very heterogeneous, and we want to understand more details. Drilling project can be very important because it will provide us with information about permeability, stress level, amount of water, heterogeneity, pseudotachylites, and all kinds of things. And we need that kind of information.
Since I'm running out of time, I'll wrap up my talk with a single figure.
